FDA releases Advisory Committee briefing documents for MannKind's Afrezza – pulmonary safety, integrity of efficacy data biggest discussion points – March 30, 2014

Executive Highlights

On Friday, the FDA released briefing documents for Tuesday’s Advisory Committee meeting for MannKind’s inhalable insulin Afrezza. Click here to view the materials. On the whole, there will be plenty of important issues for the committee to consider and debate, including Afrezza’s differentiation from subcutaneous insulin, the integrity of glycemic efficacy data, unique PK/PD characteristics, and pulmonary safety. Although there are no single factors that seem like deal-breakers on their own, some red flags (cancer imbalances, “missing data”) do show up in the documents, and we wonder how panelists will view the overall benefit/risk, particularly given that panelists aren’t always the most well prepared.

The first two draft discussion questions address Afrezza’s relative glycemic efficacy in type 1 and type 2 diabetes – both acknowledge Afrezza’s lack of A1c superiority relative to subcutaneous insulin. We expect the discussion around the efficacy data to quite nuanced, as there are so many questions to address (see below) – indeed, the two efficacy questions have a total of ten bullet points to stimulate discussion. Overall, other factors will likely drive the benefit side of the equation (convenience elements that lead to better adherence, hypoglycemia, weight, PK/PD). We were glad to see a sub-bullet that acknowledges “the importance of having an alternative route of insulin administration available” – in our view, that is a very key advantage of Afrezza given that so many patients – and HCPs! – appear to have difficulty with traditional insulin. The efficacy debate is grounded in Afrezza’s turbulent regulatory history, which includes a 2010 CRL related to a lack of data on clinical efficacy.

Pulmonary safety will once again be the major safety discussion point, given: i) a slight, non-significant numerical imbalance in lung cancer with Afrezza; ii) a slight and transient worsening of pulmonary function with Afrezza; iii) an increased incidence of cough and discontinuations due to cough; and iv) bronchial spasms in patients with pre-existing pulmonary disease. Given that Afrezza can easily be contraindicated in populations like this, we would not expect this to be a major problem; with all the chatter about the importance of individualization of therapy, we assume that the panel will not think this insulin has to work for all patient sub-groups, but one never knows. While it seems like most of these issues are either not highly clinically meaningful or are well-characterized, pulmonologists are very well-represented on the panel roster – of the 11 MDs on the panel, the plurality (five) are focused on pulmonary medicine vs. just four endocrinologists, which was surprising. As a result, we expect to see plenty of discussion on these and other pulmonary safety issues.

The FDA’s documents reference a few worrying trial execution factors, including missing data due to trial dropouts and insufficient optimization of insulin therapy in the comparator arm of the phase 3 trial in type 1 diabetes. The rest of the documents did not provide much specificity or clarity on these issues.

The advisory panel is 15 members in size, including six who have no voting history on obesity or diabetes agents in the past five years. Among the nine panelists who do have a voting history, seven have voting records that have tended to be more rather than less positive, one has a neutral record, and one has a more negative voting history (this is just information – we are not trying to generalize too much since obviously what the panels are voting on drives so much of the question). Should the advisory committee deem that the benefits do not outweigh the risks for the proposed general diabetes indication overall, we assume the panel will consider a narrower approval window for patients who might benefit the most (e.g., those struggling with post-prandial hyperglycemia and hypoglycemia). We would have assumed that MannKind would have put together an NDA resubmission designed specifically to address the FDA’s previous concerns, and we would have been surprised if the company had not expected discussion on Tuesday. See below for a table with some of the key discussion points, followed by much more background on the efficacy data, safety concerns, and panelist history.

Table 1: Possible issues to be discussed at Advisory Committee meeting, based on information included in briefing documents

Safety Concerns

  • Pulmonary function:
  • Lung cancer
  • Transient reductions in pulmonary function
  • Cough
  • Bronchial spasms in patients with asthma

Efficacy Data               

  • Relative A1c-lowering efficacy in type 1 and type 2 diabetes:
  • Reasons for more modest efficacy compared to rapid-acting analogs; is the difference meaningful?
  • Should Afrezza be compared to insulin or oral drugs in type 2 diabetes?
  • Reduced hypoglycemia compared to rapid-acting insulin
  • Lower fasting plasma glucose with Afrezza

Non-Glycemic Benefits

  • Less weight gain
  • Convenience -> adherence benefit?

Study execution

  • Possible “missing data”
  • “Imbalances” in study discontinuation
  • Was treatment titrated properly in comparator arms?


  • Unique clinical characteristics
  • Atypical dose response relationship
  • Novel PK/PD considerations due to ultra-rapid onset
  • Possible effects of variability in inhalation speed
  • “Baggage” from previous review cycles
  • “Baggage” from Exubera (Pfizer’s withdrawn inhaled insulin)
  • Is there room for conditional / subgroup-specific approval (if not overall approval)?



  • MannKind’s inhalable insulin Afrezza (Technosphere insulin inhalation system) is designed to improve both glycemic efficacy and the patient experience in both type 1 and type 2 diabetes. The product consists of an inhalation device, which administers a powder-based recombinant insulin compound. During the administration process, the patient inhales the insulin compound, which is then absorbed into the bloodstream directly through the epithelium in the lungs. This highly direct delivery pathway means that Afrezza can achieve a much more rapid onset than currently available insulins (a 12-15 minute peak vs. 60-90 minutes with rapid-acting analogs). The product’s pharmacokinetic profile is also defined by a shorter duration of action (smaller “tail”), which theoretically should reduce postprandial hypoglycemia – this effect has been borne out in clinical testing (see below). MannKind’s briefing documents state that Afrezza has the potential to reduce the “clinical inertia” that prevents type 2 diabetes patients from escalating to insulin therapy, can improve adherence in both type 1 and type 2 diabetes, and can reduce the incidence of hypoglycemia, among other benefits.
  • Despite its promising benefits, Afrezza has had a turbulent regulatory history, involving two FDA Complete Response Letters (CRLs):
    • The original FDA New Drug Application for Afrezza was filed in March 2009. The data used to make this submission used the first-generation MedTone inhaler. MannKind received a CRL in early 2010 in response to a perceived lack of data on the clinical efficacy of the candidate.
    • MannKind filed a second NDA in 2010, supported by data using a new Gen2 inhaler and the same Technosphere insulin molecule. The Gen2 inhaler was designed to provide equivalent administration of the drug compared to the MedTone inhaler, but with improved ease of use for patients. In a second CRL, the FDA requested two new phase 3 studies, one in type 1 diabetes patients, and one in type 2 diabetes patients. These trials also had to study the MedTone and Gen2 inhalers together, to confirm their clinical equivalence. The most recent NDA resubmission addresses these concerns directly.
  • We can’t help but wonder if this iteration of the review process might be affected by lingering concerns or baggage from the previous cycles. As expected, both the FDA and MannKind briefing documents provide an outline of the past review cycles. The introduction to the FDA document notes that Afrezza has a “complex regulatory history,” and that panel members should “familiarize themselves with the regulatory history” of the candidate. We assume that MannKind has built upon the previous CRLs and worked closely and appropriately with the FDA to put together the optimal application package this time around.

Glycemic Efficacy

  • We anticipate that Afrezza’s glycemic efficacy will be a key area of discussion for the panel, given that some of MannKind’s phase 3 studies have shown that the candidate provides a slightly less robust mean A1c reduction than comparator insulins. The first two discussion questions on the agenda prompt panelists to discuss this efficacy data in some detail, with a combined ten bullet points to fuel discussion. Particularly given Afrezza’s first CRL, we expect the question of efficacy will receive ample discussion.
  • Overall, we feel that Afrezza’s overall profile (convenience, less hypoglycemia, weight benefit over rapid-acting analogs) could make up for any difference in A1c-lowering efficacy vs. subcutaneous rapid-acting analogs. Perhaps more importantly, based on the most recent phase 3 data in type 1 diabetes, it is very possible that Afrezza’s effect on A1c was blunted by a reduction in hypoglycemia (see below). This may be an opportunity for the FDA to acknowledge that A1c is an imperfect glycemic measure. While of course convenience isn’t an “end” in and of itself, we think convenience is very important since it is likely to lead to better adherence from the patient in our view; to boot, it also may well be much better for HCPs, many of whom are struggling when it comes to prescribing insulin at the right time, etc.
  • The most recent data on Afrezza’s glycemic efficacy comes from MannKind’s phase 3 studies reported last August – one in type 1 diabetes, and the second in type 2 diabetes. The two trials were initiated as a direct outcome of the second FDA CRL for Afrezza, and were designed in cooperation with the FDA to specifically address outstanding agency concerns about the product. Below, we summarize the key findings of these trials – read our full report on the results for more details.
    • The trial in type 1 diabetes found that Afrezza provided non-inferior A1c reductions relative to insulin aspart (-0.2% and -0.4%, respectively). The 95% confidence interval for the between-group difference (0.02%-0.36%) just narrowly achieved the predetermined non-inferiority threshold of 0.40%. However, when examined through a different statistical lens, the mean A1c reduction with Afrezza was significantly less than the reduction seen with insulin aspart (by a point estimate of 0.22%). Additionally, fewer patients in the Afrezza group (Gen2 inhaler group) achieved an A1c below 7.0% (18% vs. 31% with insulin aspart; p = 0.02) or below 6.5% (8% vs. 13% with insulin aspart; p=0.2) at week 24. We view the non-inferiority data as the most important, as the study was designed as a non-inferiority study. However, the statistical review section of the FDA briefing documents states that “the comparative efficacy shown here [in Trial 171] was not compelling” due to the upper bound of the 95% CI coming so close to 0.40%.
      • Notably, Afrezza was associated with better fasting plasma glucose, less weight gain, and less hypoglycemia (including severe hypoglycemia) than insulin aspart. There were 9.8 hypoglycemic events per subject-month with Afrezza vs. 14 events with insulin aspart (p<0.0001), and eight events of severe hypoglycemia per 100 subject-months with Afrezza vs. 14 with insulin aspart (p=0.10, not significant). Mean change in fasting plasma glucose was -25 mg/dl with Afrezza vs. +10 mg/dl in the insulin aspart group (p=0.0027). Patients on Afrezza lost, on average, 0.4 kg (1 lb) relative to the 0.9 kg (2 lbs) gained with insulin aspart (p=0.01). These results could suggest that Afrezza’s apparent A1c-lowering disadvantage in this trial may be the result of reduced hypoglycemia.
      • Study design: The study included 174 patients randomized to Afrezza and 170 randomized to the comparator arm – the study also included an arm of patients using Afrezza with the older MedTone inhaler, to examine the equivalence of that device with the newer Gen2 device. The results are broadly seen as confirming the devices’ comparability, allowing data from older trials with the MedTone inhaler to be used to support Afrezza’s safety and efficacy.
    • The trial in type 2 diabetes (Trial 175) found that Afrezza plus orals led to a significantly greater A1c reduction relative to oral antidiabetic therapy alone (-0.8% vs. -0.4%, respectively; p<0.0001; baseline A1c ~8.0%). However, the incidence of mild/moderate hypoglycemia was twice as high in the Afrezza + orals group vs. in the oral + placebo powder group (67% vs. 30%, p<0.0001).
      • Study design: Type 2 diabetes patients on oral antidiabetic therapies were randomized to continue those therapies along with a placebo powder inhalation (n=176) or to add Afrezza (Gen2 inhaler) to their oral therapies (n=177). The study ran for 24 weeks. Notably, patients were not permitted to adjust the doses of their oral medications during the study without discussion between the principal investigator and medical monitor, a design element that could conceivably explain some of the increased hypoglycemia seen in the Afrezza arm.
    • Data from the studies that supported MannKind’s previous applications for Afrezza were included in this most recent submission. As background, the previous phase 3 trial in type 1 diabetes barely missed the non-inferiority margin against injectable insulin. Previous trials also showed an increase in diabetic ketoacidosis with Afrezza, which could have been correlated with the slight difference in glucose-lowering. No further DKA events have been seen in the most recent studies, but the reviewer comments in the FDA briefing documents still characterizes the past DKA imbalance as potentially concerning.
    • We found it curious that in the FDA’s discussion question on Afrezza’s efficacy in type 2 diabetes, the agency focused on the comparison against subcutaneous insulin. By contrast, the most recent phase 3 study in type 2 diabetes compared Afrezza plus orals to orals alone. We wonder how the panel will weigh the different comparator drugs, particularly as it relates to a type 2 diabetes indication.
  • It is possible that the slightly reduced A1c-lowering efficacy seen with Afrezza relative to injectable rapid-acting insulin is result of reduced hypoglycemia. We certainly hope this issue is raised and that panelists look beyond just A1c – time in zone, fasting plasma glucose, and weight.
  • We feel that the phase 3 data on Afrezza is an excellent example of the pitfalls of relying too heavily on A1c, and not enough on measures such as “time in zone” – admittedly, “time in zone” have only recently become easier to measure such as with Dexcom’s G4, where accuracy in hypoglycemic ranges aren’t problematic. Big picture, it would be unfortunate if a slight difference in a surrogate measure (A1c) is used as a rationale to not approve a drug, if that difference is due to actual improvements in diabetes management overall. Although the briefing documents and much of the discussion questions focus on A1c, the sub-bullets in the specific discussion questions on efficacy seem to leave some room for a nuanced discussion of metrics beyond A1c.
  • If the panel does not believe that Afrezza’s efficacy and other benefits support a full approval for type 1 and type 2 diabetes, we assume a conditional approval will be considered, in patients with poor postprandial control (either hyperglycemia or hypoglycemia) as one example, or patients with a major fear of injections, etc. Both efficacy questions on the agenda contain a sub-bullet on whether Afrezza might be an appropriate substitute for subcutaneous mealtime insulin in a “specific subgroup of individuals,” if not in all individuals with type 1 and/or type 2 diabetes. Given the uncertainty surrounding any new drug approval, conditional approvals seem like a new “minimum” to expect to get new therapies into the hands of patients that need them the most. Certainly, the blanket-approval, “one-strike-you’re-out” pattern that has emerged as of late (e.g., Novo Nordisk’s insulin degludec) has already had repercussions on investment and drug development.

Trial Design and Data Issues

  • The briefing document outlines a series of potential issues related to study execution – factors include missing data and imbalances in study dropout rates. We see these topics as relatively notable, as these concerns have the potential to undermine the real or perceived integrity of MannKind’s data, and therefore the chances of Afrezza’s approval. Of course, we cannot imagine that the FDA could raise many doubts about the original plan and design of the two new pivotal phase 3 trials, given that they were specifically designed in cooperation with the FDA to address the requests from the last CRL.
    • The FDA briefing document cites the handling of “missing data” from the two pivotal phase 3 studies (data from patients lost to follow up) as a potential issue hindering the interpretation of those studies’ results. “Missing data and assumptions made in the handling of missing data” is included in a list of draft discussion points in the introduction of the FDA briefing documents. Most sensitivity analyses showed that the missing data would likely not have affected the primary results, although in the pivotal type 1 diabetes study (study 171), when an A1c rise of 0.4% was applied to every patient who discontinued in the Afrezza group, this pushed the 95% CI beyond the bounds of non-inferiority. In the type 2 diabetes trial (study 175), all sensitivity analyses showed that the missing data should not have affected the primary results.
    • In both trials, there were differences in the dropout rates between the Afrezza and comparator arms. In study 171 (type 1 diabetes), 25% of the Afrezza arm dropped out prior to completion, relative to only 11% of the insulin aspart arm. However, in study 175 (type 2 diabetes), the dropout rate was higher in the placebo group than in the Afrezza group (30% and 21%, respectively). Differential drop out rates may produce difficulties in interpreting the effect of missing data. We wonder if the positive dropout data in type 2 diabetes will be linked to a particular adherence advantage with Afrezza in this population.
    • The FDA briefing document also takes issue with the optimization of insulin in the control arm of the type 1 diabetes study (study 171). This suggests that if the control arm was artificially less tightly controlled than it would have otherwise been, then Afrezza’s efficacy may have been overestimated.

Pulmonary Safety

  • As with previous FDA reviews of Afrezza, pulmonary safety will likely be the most significant family of safety issues for the panel to discuss – especially given that there are several pulmonologists on the panel (see our panel summary below). However, given that the pulmonary impacts of Afrezza have generally been characterized and appropriate label restrictions have been proposed (including contraindications for asthma and COPD), we do not see pulmonary safety as a primary reason that Afrezza would not be approved. The pulmonary safety issues that will likely be discussed fall into four main groups: (i) a slight, non-significant imbalance in lung cancer seen in phase 3; (ii) an imbalance in the incidence of cough and discontinuation due to cough; (iii) a modest, transient, and relatively well-characterized reduction in lung function seen with Afrezza; and (iv) an increase in bronchospasm in patient with underlying lung disease.
    • The slight imbalance in lung cancer cases seen in pooled safety data on Afrezza does not seem particularly worrying. There were four cases of lung cancer reported in patients exposed to Afrezza, compared with none in the comparator arm (difference not statistically significant). The incidence rate in the Afrezza pooled cohort is comparable to what would be expected in the general population, and as a whole the data do not support the idea that this should be a decision-altering concern for the panel. However, Afrezza does still carry some of the baggage from Pfizer’s inhaled insulin Exubera, which was associated with an increase in lung cancer. The results from Exubera’s FUSE study (which detailed the increase in lung cancer) were included in the FDA briefing documents, and lung cancer is listed as one of seven discussion questions on the meeting agenda. We believe that any doubt on this issue could be addressed through a post-marketing study. Indeed, the FDA’s briefing packet urges postmarketing follow-up on pulmonary malignancies should Afrezza be approved, either using a registry with methodology to reduce detection bias or a large randomized controlled trial.
    • An imbalance was seen in both the rate of cough and the rate of discontinuation due to cough. Cough was the most frequent treatment-related adverse event in the pooled safety population (type 1 and type 2 diabetes), affecting 27% of patients on Afrezza, 20% of patients on placebo inhalation powder, and 5% of non-inhalation comparator patients. According to the MannKind briefing documents, there was a ~21% excess incidence of cough versus comparators in the first week of Afrezza use, declining to only ~2% by the third month of treatment. Cough was generally mild, transient, and occurred within 10 minutes of administration. However, the fact that there was also an imbalance in trial withdrawal due to cough means that this is not an effect than can be seen as minor or clinically insignificant. Education and increased experience with the device could help patients acclimate to the increase in cough, which did not seem to be tied to any further adverse pulmonary outcomes.
    • We do not anticipate that the decline in lung function seen with Afrezza will be a very controversial factor. A modest (0.4-0.6 L) decline in FEV1, a metric of pulmonary function, was seen in the first three months of therapy with Afrezza; the difference did not grow over the time period, and the effect has been shown to disappear following cessation of therapy. Given that the FDA would likely contraindicate Afrezza in patients with asthma, COPD, or other chronic lung disease, we do not see this as a major issue for most patients. We do wonder if this slight effect will dissuade athletes from considering Afrezza, but that certainly shouldn’t be seen as a factor that should get in the way of approval.
    • “Bronchospasm” (sudden constriction of bronchial walls) may come up during the Advisory Committee meeting as a pulmonary side effect, but as with pulmonary function, we wouldn’t expect it to be a very major issue. Although patients with pre-existing lung disease were not enrolled in phase 2/3 clinical trials for Afrezza, smaller safety studies were conducted in patients with asthma and COPD. Those studies showed severe adverse events of bronchospasm in patients with asthma. Given the proposed contraindication for asthmatic patients (among others), we do not see this as a major issue for patients, although it could be an issue for patients with undiagnosed or emergent asthma.

Clinical Pharmacology

  • Given how novel an approach Afrezza represents, as well as its very rapid onset of action, perhaps it is not surprising that the briefing documents list some areas of uncertainty with regards to some of Afrezza’s clinical characteristics. The introduction of the FDA briefing documents lists four specific points in this area:
    • “The novel pharmacokinetic and pharmacodynamic profile of Afrezza as it relates to meal time coverage.” As we have heard from MannKind and other KOLs previously, Afrezza’s onset of action is so rapid that it changes dosing and timing. Afrezza can be administered immediately before (or even shortly after) the initiation of a meal vs. the recommendation for ~30 minutes prior to the meal with current rapid-acting analogs. Afrezza’s shorter tail also seems to produce less post-meal hypoglycemia. The implications of this rapid action for patients and providers will likely come up in discussion on Tuesday.
    • “Lower bioavailability compared to subcutaneous insulin.” Afrezza has ~20-30% of the bioavailability of subcutaneous insulin, which MannKind has accounted for using a dose conversion algorithm (10 units of Afrezza per four units of subcutaneous insulin). Afrezza has been submitted to the FDA with two cartridge doses: an equivalent to three-units of a rapid-acting analog and an equivalent to six units. Nine and 12-unit dose equivalents are expected to be filed following FDA approval.
    • “Atypical dose response relationship.” In earlier clinical studies, increasing doses of Afrezza demonstrated a less-than-linear effect on glucose lowering above 60 units (~20-24 units of subcutaneous insulin). For people with type 2 diabetes who are very insulin resistant and require large meal boluses, this could cause issues with patients not receiving enough insulin at high doses.
    • “Specific issues related to method of drug delivery.” The briefing document states that the specific manner in which a patient inhales the insulin from the inhaler can have an effect on variables such as particle size. The clinical significance (if any) of this variability is not fully clear.

Meeting Agenda

  • Meeting agenda: At 8 AM (ET) the meeting will begin with an introduction of committee by Dr. Robert Smith (Alpert Medical School of Brown University, Providence, RI), followed by a conflict of interest statement by Karen Abraham-Burell (FDA, Silver Spring, MD) and introductory remarks by Dr. Jean-March Guttier (FDA, Silver Spring, MD). From 8:35 to 10:05 AM. MannKind will deliver its presentations. These are to include an introduction by John Bedard (SVP, Regulatory Affairs, MannKind), and discussions on the medical need by Dr. Janet McGill (Washington University School of Medicine in St. Louis, MO), Afrezza’s pharmacology by Dr. Robert Baughman (SVP, Clinical Sciences, MannKind), and Afrezza’s clinical efficacy by Dr. David Bregman (Medical Director, Clinical Development, MannKind). Additionally, both Dr. Nikhil Amin (VP, Clinical Development, MannKind) and Dr. John Gerich (University of Rochester, School of Medicine, NY) will speak to Afrezza’s clinical safety. MannKind’s closing remarks will be made by Dr. Baughman. The FDA’s presentations consist of an overview of Afrezza’s pharmacology, clinical efficacy, pulmonary and non-pulmonary safety, and epidemiology and post-market surveillance. We think the inclusion of an FDA presentation on this last topic is notable (though not surprising), and will pay close attention to how appropriate the Agency characterizes post-market options.

Panel Voting History

  • Panelists’ voting histories from previous Advisory Committee meetings and areas of expertise suggest that much of the conversation is going to focus on Afrezza’s pulmonary safety. Indeed, based on our research into people’s training, of the 11 MDs on the panel, the plurality (five) are focused on pulmonary medicine. Two of these pulmonologists are experts in lung cancers (one in the prevention of lung cancer). For comparison, four of the MDs are endocrinologists, one is an obesity specialist, and one is a cardiologist. Among the nine panelists who have voted on the approval of an obesity or diabetes agent in the past five years, seven have voting records favoring approval, one has a neutral record, and one has a history of voting against approval. We are particularly encouraged to see that Dr. Abraham Thomas (Henry Ford Hospital, Detroit, MI) and patient representative Ms. Rebecca Killion (Washington, DC) are on this panel. We think that Dr. Thomas is an eloquent, sharp, and extremely well-respected panel member, as is Ms. Killion, who consistently advocates for increasing the number options available to patients with diabetes.

Panelist Name


Overall (Y:N) Vote History for EMDAC Diabetes/Obesity Drug Approval Meetings in Past Five Years***

Level of Influence**

Key Voting Considerations*

Erica Brittain




- Data quality

- CV risk

William Calhoun


Has not served on an EMDAC diabetes/obesity drug approval panel in the past five years

David Cooke

Pediatric Endocrinologist



- Unmet need

- R&D incentives

Katherine Flegal



Moderate-High (former JAMA editor)

- Drug safety monitoring/population health

- Efficacy and niche roles

Diana Hallare

Consumer Representative



- Cancer risk

- CV risk

Ed Hendricks



Moderate - High

- Unmet need

- R&D incentives

Rebecca Killion

Patient Representative



- Unmet need

- Quality of life

Morris Schambelan


Has not served on an EMDAC diabetes/obesity drug approval panel in the past five years

Robert Smith




- CV risk

- “Do no harm”

James Stoller


Has not served on an EMDAC diabetes/obesity drug approval panel in the past five years

Erik Swenson


Has not served on an EMDAC diabetes/obesity drug approval panel in the past five years

Eva Szabo

Pulmonary Oncologist

Has not served on an EMDAC diabetes/obesity drug approval panel in the past five years

Abraham Thomas



Very High

- Data quality

- CV risk

Peter Wilson




- CV risk

- Cancer risk

Antoinette Wozniak

Thoracic Oncologist

 Has not served on an EMDAC diabetes/obesity drug approval panel in the past five years

*Key voting considerations: Our impression of the issue(s) that weigh the heaviest on a panelist’s vote.  **Level of influence: How much a panelist’s opinion appears to sway the votes of fellow members. ***Overall voting history: The count includes the results of votes for/against drug approvals at FDA EMDAC meetings since 2008. This includes the following drugs: Orexigen’s Contrave (bupropion/naltrexone), J&J’s Invokana (canagliflozin), AZ’s Farxiga (dapagliflozin – 2013 and 2011), Novo Nordisk’s Tresiba (insulin degludec), Novo Nordisk’s Victoza (liraglutide), Arena’s Belviq (lorcaserin – 2010 and 2012), Vivus’ Qsymia (phentermine/topiramate ER – 2010 and 2012), and AZ’s Onglyza (saxagliptin) [1] The following diabetes- and obesity-related EMDAC meetings were not included in this table because panelists were not asked to vote on a drug’s approval: Abbott’s Meridia (sibutramine) market removal, GSK’s Avandia (rosiglitazone) CV safety analyses (2010 and 2013), the 2012 meeting on Obesity CV risk assessment, and the 2008 meeting on Diabetes Drug CV risk assessment.

  • (Acting Chair) Robert Smith, MD (Endocrinologist, Alpert Medical School of Brown University, Providence, RI): Overall, Dr. Robert Smith has been a moderate voter. He voted against the approval of Novo Nordisk's Tresiba (insulin degludec) at the FDA Advisory Committee meeting in November 2012, characterizing the decision as being very difficult. He ultimately decided to vote "no" because the magnitude of the benefit did not outweigh the cardiovascular risk signal. He was further moved to vote against Tresiba’s approval given the availability of reasonably viable treatment alternatives – a consideration that could negatively impact his thinking on MannKind’s Afrezza, depending on how differentiated he sees the agent from other fast-acting insulins. In contrast, in the 2013 Advisory Committee meeting for AZ’s Farxiga (dapagliflozin), Dr. Smith voted for the agent’s approval emphasizing the importance of post-marketing studies, an openness that we think could favor Afrezza’s approval. In June 2013 Dr. Smith voted to continue the REMS/ETAUS for GSK's Avandia (rosiglitazone) because of lingering concerns over the drug's cardiovascular safety. His voting history on anti-obesity medications is more positive and includes "yes" votes on Arena's Belviq (lorcaserin) and Vivus' Qsymia (phentermine/topiramate).
  • Erica Brittain, PhD (Statistician, NIAID, NIH, Bethesda, MD): Unsurprisingly given Dr. Erica Brittain’s expertise in biostatistics, data is the number one voting factor for her. In cases of uncertainty, Dr. Brittain tends to prefer gathering more data vs. approving the drug and collecting data post-approval. For example, she voted against the approval of Novo Nordisk’s insulin degludec (Tresiba), citing the CV signal being too great to be addressed post-approval. However, she voted in favor of approving J&J’s canagliflozin, citing the drug’s robust efficacy (perhaps even superiority) and “fairly promising” cardiovascular profile. At the second Advisory Committee meeting for Arena's Belviq, she voted in favor of approving the drug, though was not happy with the abundance of missing data – she felt this compromised the integrity of randomized comparisons of efficacy and safety.
  • William Calhoun, MD (Pulmonologist, University of Texas Medical Branch, Galveston, TX): Dr. William Calhoun has not previously served on an EMDAC committee concerning a diabetes or obesity drug in the last five years. As background, Dr. Calhoun is the Vice Chair for Research in the University of Texas Medical Branch’s Department of Internal Medicine. His research interest is focused on asthma.
  • David Cooke, MD (Pediatric Endocrinologist, Johns Hopkins University School of Medicine, Baltimore, MD): Dr. David Cooke appears to make a balanced assessment of each agent he has considered. Dr. Cooke thought that Tresiba’s (Novo Nordisk’s insulin degludec) pharmacokinetics would be enough of an advancement for people with type 1 and type 2 diabetes to justify approving the agent despite its “far from proven” cardiovascular risk. Similarly, he voted in favor of J&J’s Invokana (canagliflozin), citing its efficacy and “sufficiently reassuring” safety data. Though he noted that Invokana’s safety data was incomplete, he warned against placing an “excessive burden” on the development of such agents, which a “very important and needy population” requires.
  • (Consumer Representative) Diana Hallare, MPH (Visalia, CA): Ms. Diana Hallare joined the FDA’s Endocrinologic and Metabolic Drugs Advisory Committee (EMDAC) as a voting member on October 2, 2013. At the Advisory Committee for AZ’s Farxiga (dapagliflozin), she voted against the candidate’s approval on the grounds that there was a potential risk for cancer, infections, and increased CV risk. Ms. Hallare is not a vocal committee member and we don’t believe she has been articulate enough to sway the panel in the past on most issues.
  • Ed Hendricks, MD (Obesity Specialist, Center for Weight Management, Roseville, CA): As a practicing clinician treating obese patients, Dr. Ed Hendricks clearly recognizes the unmet market needs of patients (particularly people with obesity) and has been one of the most liberal voters on the panel. Dr. Hendricks voted for the approval of Novo Nordisk’s Tresiba (insulin degludec), since the thought the agent would achieve one of the holy grails of diabetes management: truly once-daily insulin injections – it will be interesting to hear how Dr. Hendricks characterizes Afrezza’s lack of injections. In 2011, Dr. Hendricks voted in favor of AZ’s Farxiga (dapagliflozin) and praised the drug for its effects on weight loss and its insulin-independent mechanism of action; he also commended its potential ease of use as an oral medication. While he suggested safety issues were of some concern, he was satisfied that a “post-marketing study would settle some of those issues.” He has also acknowledged that there is a limit to what can be learned from clinical trials and advocated to the agency that understanding that limit was critical to “taking medicine forward and introducing innovative things.”
  • Katherine Flegal, PhD (Senior Scientist and Distinguished Consultant, CDC, Hyattsville, MD). Dr. Katherine Flegal has a split 3:3 voting record in diabetes and obesity-related advisory committee meetings over the past five years. She voted for the approval of Arena/Eisai’s Belviq (lorcaserin) and Vivus’s Qsymia (phentermine/topiramate) at both candidates’ second Advisory Committee meetings in 2012, while she had voted against their approvals at their respective first committee meetings in 2010. During both sets of meetings she expressed concern about the agents’ unclear efficacy. What appeared to drive her change in voting status was greater clarification of the safety risks of both agents, as well as the promise of post-approval trials and monitoring. Based on Dr. Flegal’s focus on Belviq and Qsymia’s efficacy, it is likely she will focus on the FDA’s question about the efficacy of Afrezza. However, her willingness to approve the anti-obesity agents despite her efficacy concerns, suggests that if she feels risk has been adequately mitigated she may remain open to approval.
  • (Patient Representative) Rebecca Killion (Washington, DC): Ms. Killion is quite well respected in the regulatory community and it is our impression that virtually all panel members in the past have trusted her astute perspective on drug candidates. As a patient representative, Ms. Rebecca Killion consistently advocates for increasing the number of quality drug options available to patients with diabetes. Overall, her comments can often be distilled to “stay safe, stay sane.” She voted in favor of Tresiba’s (Novo Nordisk’s insulin degludec) approval, since she believed it represented an important step forward for basal insulin therapy and we suspect she also did not want to slow down progress toward next-generation candidates that may be more breakthrough in nature. She thought the benefits outweighed the risks and was not convinced that the cardiovascular safety signal was real. Similarly, Ms. Killion viewed J&J’s canagliflozin as “very encouraging,” since it represented a new mechanism of action, and improved A1c while addressing concerns about weight and hypoglycemia. Her central concern with the agent was if patient adherence would be negatively impacted by the prevalence of urinary tract infections (UTIs). In this vein, we think Ms. Killion may pay close attention to the negative impact the cough associated with Afrezza could have on patient adherence, but that she will balance this with the quality of life improvement patients with needle phobias might experience having access to this agent.
  • Morris Schambelan, MD (Endocrinologist, University of California, San Francisco, CA): Dr. Morris Schambelan has not served on an EMDAC panel for a diabetes or obesity drug over the last five years. He has diverse research interests, ranging from HIV-related lipodystrophy to the impact of gastric bypass surgery on changes in glycemic control and pancreatic hormones.
  • James Stoller, MD, MS (Pulmonologist, Cleveland Clinic, Cleveland, OH): Dr. James Stoller has not served on an EMDAC panel for a diabetes or obesity drug over the last five years. Dr. Stoller’s specialty interest is in alpha-1 antitrypsin deficiency (a condition in which the body does not make enough of a protein that protects the lungs from damage, which can lead to emphysema) and chronic obstructive pulmonary disease (COPD). Additionally, Dr. Stoller is the Executive Director of Physician Leadership Development at the Cleveland Clinic and is Chairman of the Cleveland Clinic’s Education Institute, which oversees many training programs for HCPs and healthcare executives.
  • Erik Swenson, MD (Pulmonologist, University of Washington, Seattle, CA): Dr. Erick Swenson has not served on an EMDAC panel for a diabetes or obesity drug over the last five years. In addition to his training in pulmonology, Dr. Swenson also completed a fellowship in pharmacology. As a result, it appears that his research interests tend to center on the influence of different chemical compounds on lung volume and function – a key question given the association of Afrezza with a mild, transient decline in lung function.
  • Eva Szabo, MD (Pulmonary Oncologist, National Cancer Institute, Rockville, MD): Dr. Eva Szabo has not served on an EMDAC panel for a diabetes or obesity drug over the last five years. She is Chief of the Lung and Upper Aerodigestive Cancer Research Group at the National Cancer Institute (NCI). Notably, this group is located in the NCI’s Division of Cancer Prevention, and she oversees lung, head, and neck cancer prevention trials.
  • Abraham Thomas, MD, MPH (Endocrinologist, Henry Ford Hospital, Detroit, MI): Dr. Abraham Thomas is an eloquent, sharp, and extremely well respected panel member. In his former role as EMDAC chairperson, Dr. Thomas skillfully synthesized panelist discussions on questions the FDA posed to them practically real-time. Dr. Thomas’ overall liberal stance generally favors drug approvals with promises to collect data post-approval. In 2013 and 2011, Dr. Thomas voted for the approval of AZ’s Farxiga (dapagliflozin). In 2013, he was reassured about CV safety given that the agents CVOT had already begun. In 2011, he noted the unrealistic nature of assessing cancer risk in a pre-approval setting – an opinion that could benefit Afrezza’s approval, given the slight imbalance in cancer events seen in its development. As another example, Dr. Thomas also voted to approve J&J’s canagliflozin, noting his general comfort with the data, and confidence that more was coming.
  • Peter Wilson, MD (Cardiologist, Emory University School of Medicine, Atlanta, GA): At the FDA Advisory Committee meeting for AZ’s Farxiga (dapagliflozin), Dr. Wilson voted for the agent’s approval expressing confidence the agent’s CVOT (DECLARE) will resolve the candidate’s cardiovascular risk issue. Dr. Wilson, however, held that breast cancer risk was still a concern for the agent, as it might take some time to manifest.
  • Antoinette Wozniak, MD (Pulmonary Oncologist, Karmanos Cancer Institute, Detroit, MI): Dr. Antoinette Wozniak has not served on an EMDAC panel for a diabetes or obesity drug over the last five years. As background, Dr. Wozniak’s clinical interest is in the treatment of lung cancer, mesothelioma, and thymus gland cancer. Her research interests fall in the multidisciplinary care of people with lung cancer, as well as the development of new drugs for the treatment of lung cancer and mesothelioma.

Appendix A: Draft Questions

  1. DISCUSSION: Trials in type 1 diabetes demonstrate that Afrezza provides numerically and  statistically less HbA1c reduction than a comparator subcutaneous insulin. Discuss whether the applicant has demonstrated that Afrezza is an effective treatment for patients with type 1 diabetes mellitus. In your discussion, please address each of the following issues:
    1. The impact of inadequate treatment optimization in the control arm on efficacy determination
    2. The impact of missing data on efficacy determination
    3. The relationship between Afrezza’s clinical pharmacology (e.g., extent and duration of action, dose-response, inhalation flow-rate dependence) to its effectiveness as a “prandial” insulin
    4. Your level of concern with regard to how differences in efficacy will impact disease specific risks (e.g., risk of diabetic ketoacidosis) beyond achievement of therapeutic goals (i.e., prevention of diabetes related complications)
    5. The importance of having an alternative route of insulin administration available
    6. Comment on whether Afrezza is an appropriate substitute for subcutaneously administered mealtime insulin for most patients with type 1 diabetes or for a specific subgroup of  individuals with the disease. If you believe the latter, describe this subgroup.
  2. DISCUSSION: Trials in type 2 diabetes demonstrate that Afrezza is superior to placebo but is less effective than a short acting subcutaneous insulin comparator. Discuss whether the applicant has demonstrated that Afrezza is an effective treatment for patients with type 2 diabetes mellitus.
    1. Comment on the specific clinical setting where this agent is likely to be most useful.
    2. Discuss your level of concern with regard to data suggesting a less than dose proportional glucose lowering response and its potential impact on achievement of glucose targets in this population.
    3. Comment on any other issues discussed in the context of type 1 diabetes that you view as relevant to type 2 diabetes.
    4. Comment on whether Afrezza is an appropriate substitute for subcutaneously administered mealtime insulin for most patients with type 2 diabetes or for a specific subgroup of individuals with the disease. If you believe the latter, describe this subgroup.
  3. DISCUSSION: Discuss the pulmonary safety findings in the Afrezza clinical development program (acute bronchospasm and pulmonary function decline over time).
    1. Comment as to whether the pulmonary safety data (6 months with Gen2 device and 2 years with MedTone device) are sufficient to address the pulmonary safety of Afrezza.
    2. Discuss your level of concern with the pulmonary risks.
  4. DISCUSSION: Discuss your level of concern with regard to the possible lung cancer risk with Afrezza use.
  5. DISCUSSION: Discuss any other risk(s), which were not covered above.
  6. VOTE: Based on data in both the briefing materials and presented at today’s meeting, has the applicant demonstrated that Afrezza is safe and effective for the treatment of adult patients with type 1 diabetes mellitus to support approval?
    1. DISCUSSION: If yes, please explain your rationale. Provide any recommendations you might have for post-marketing studies to evaluate identified safety signals.
    2. DISCUSSION: If no, please explain your rationale. If appropriate, what further data should be obtained?
  7. VOTE: Based on data in both the briefing materials and presented at today’s meeting, has the applicant demonstrated that Afrezza is safe and effective for the treatment of adult patients with type 2 diabetes mellitus to support approval?
    1. DISCUSSION: If yes, please explain your rationale. Provide any recommendations you might have for post-marketing studies to evaluate identified safety signals.
    2. DISCUSSION: If no, please explain your rationale. If appropriate, what further data should be obtained?


-- by Manu Venkat, Hannah Deming, Jessica Dong, Adam Brown, and Kelly Close


[1] Because the pivotal voting questions for liraglutide and saxagliptin were not straightforward yes/no approval questions, Close Concerns interpreted responses in the following manner: for both of these agents, panelists were asked to vote “yes” or “no” on whether the data on specific safety concern(s) supported approvability (liraglutide: regarding thyroid C-cell tumors and papillary thyroid cancer; saxagliptin: regarding unacceptable excess CV risk). We inferred that if a member voted “yes” in favor of safety to both liraglutide’s voting questions 3 and 4 then he/she was expressing support of the drug’s approval. If a member voted “no” to at least one of the two questions, we assumed he/she was expressing opposition to approval. For saxagliptin, we counted a “yes” vote to question 1 as a “yes” approval vote. A more granular breakdown of votes for liraglutide and saxagliptin can be found in Tables 4.1 and 4.2.